Damage identification in composite panels using acousto-ultrasonic waves
Operational cost in aviation is an important factor of the total cost of an aircraft. These cost can be divided in direct maintenance cost and indirect costs due to unscheduled maintenance time. The use of composites introduces the additional uncertainties: invisible defects can require unexpected ground time. By embedding sensor systems in composite aircraft structures, inspection can evolve to monitoring, known as Structural Health Monitoring. The advances in this field are on the one hand considerable, but on the other hand the TRL level of these solutions is yet insufficient to find real applications. The challenges are the complexity of the composite material, the high density of sensors required for a robust damage assessment and the high frequencies required for the interrogation of the structure. A solution with a lot of potential is the Smartlayer, originally developed by the Stanford university. A layer of sensors is embedded in the structure, where each sensor can act as an actuator while the others act as receivers. By sequentially assigning each sensor as an actuator, a series of actuator-sensor paths is formed, covering the area enclosed by the sensors. The time signal of the paths contains information on the current state of the structure. This information can be extracted in multiple ways, though in all cases, a reference measurement is required. Moreover, the way the path information is converted into a possible damage location is far from trivial. A composite panel with Omega stringers is equipped with a SMART Layer sensor system. Measurements are performed on the pristine structure and again between and after multiple impact damages were applied. During the measurements, an acousto-ultrasonic signal is send from each sensor to all other sensors in a predefined set of sensors (the workspace). A damage index surface plot for the workspace area is calculated using different algorithms, based on correlation functions and on actuator-sensor path information of the signals. Visualisation of the damage and control of the settings of the damage identification algorithms is implemented in a Matlab based software with graphical interface. It visualises the damage locations, based on user settings and also allows the user to select and isolate damaged regions for further analysis. The research has led to a better understanding of how to transform the path damage indices to a damage location, but also provides an intuitively method for the user to optimise the settings for the damage identification for each individual case. The graphical interface allows the user to extract relevant information on the structural integrity of the structure without requiring the user to have thorough understanding of all the details of the algorithms to analyse the time response data. The output of the data can be used as input for Probability of Detection models, further supporting the entire maintenance decision procedure. Future developments are directed towards further improvements of the damage assessment algorithms and increasing the level of interaction between the user and the software.